Re: db5.3 package maintainance
On Sun, Aug 12, 2018 at 11:04:42PM +0200, Karsten Merker wrote:
> [CCing debian-riscv@lists.debian.org; please keep the list
> in CC on replies]
>
> Hello Dmitrijs, hello Ondřej,
>
> I'm writing directly to you because you are listed as
> uploaders for src:db5.3 and the package maintainer address
> (pkg-db-devel@lists.alioth.debian.org) is no longer valid,
> therefore sending an email to the BTS would just result in a
> bounce.
I just realized you might have relied solely on that list, and not be
subscribed directly to the packages (unlike Maintainer, Uploaders have
to enable that manually). This means you might have lost some of BTS
mails from me. We did ping you guys on IRC repeatedly, though.
> I'm one of the riscv64 porters and db5.3 is one of the few
> remaining packages that block us from being able to debootstrap
> a riscv64 system. The package currently FTBFS on all
> architectures due to changes in the JRE baseline (cf.
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=873976).
There are two other FTBFS issues beyond what's in the BTS.
> Adam Borowski has tried to NMU the package with the one-line
> patch attached to the bugreport, but that has failed as the
> source tarball contains a pre-built MS help file without
> corresponding source which has led to a reject from FTP masters
> (cf. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=873976#38),
> so future uploads would require either removing the offending
> file from the source tarball or adding the corresponding source
> if available.
There's another NMU in DELAYED, slated for Thursday early hours.
It includes the tarball repack, FTBFS with Java 9, a different FTBFS with
Java 10, and a FTBFS with gcc-8 (the latter two became an issue only
recently). It doesn't include updating the maintainer address.
There's no working git repository I could have put the patches into, but
they're obvious -- other than the tarball repack, that is. Because of
debian/copyright not using the machine-readable format, Files-Excluded
can't be used, thus the tarball in DELAYED has been repacked manually.
> It would be great if you could take a look at the issue, or in
> case somebody else takes care of the package now, provide us with
> corresponding contact information.
If you're not satisfied with the NMU, please cancel it, or say how else I
can integrate the fixes into a repository you have somewhere.
> If you should upload a new package version, it would also be
> great if you could add riscv64 to the "JAVA_BROKEN_ARCHS" list
> (as is already the case for m68k) because we currently don't have
> a working JDK for riscv64.
It did build successfully, and IIRC this includes a testsuite. Not sure
what I'm missing.
Meow!
--
⢀⣴⠾⠻⢶⣦⠀ So a Hungarian gypsy mountainman, lumberjack by day job,
⣾⠁⢰⠒⠀⣿⡁ brigand by, uhm, hobby, invented a dish: goulash on potato
⢿⡄⠘⠷⠚⠋⠀ pancakes. Then the Polish couldn't decide which of his
⠈⠳⣄⠀⠀⠀⠀ adjectives to use for the dish's name.
Reply to: